Update: Our website is now live!
TL;DR: “Stats for Startups” is a research project that seeks to illuminate and mitigate bias in early-stage financing. A collaboration between Charge Ventures and Professor Alessandro Piazza of Rice Business School, our goal is to make better benchmarking for common startup KPIs freely available to all entrepreneurs. The project is made possible through generous support from the Ewing Marion Kauffman Foundation for Entrepreneurial Research.
Follow Charge on Medium, Linkedin, or our new but soon to be famous Twitter for updates on the project. Please reach out if you want to get involved!
What’s your (startup) worth?
“If you are lucky, this will be the only seed round you ever raise.”
I don’t remember which VC told me this, but I never forgot it. I was 27 years old, neck-deep into raising my first round of institutional venture capital for @Sonar. Frankly, I had no idea what I was doing.
It showed in my cap table.
Nine months earlier, I had raised my first $150k from my employers for a mere 80% of my company. Was that a “good” deal or a “terrible” one? It’s a matter of perspective. Was mine an angel round? Terrible deal. Was I being incubated? Par for the course.
Unfortunately, as a newbie (read: noob) to the world of venture capital, I lacked the necessary experience and network to evaluate the offer. Asking around brought opinions ranging from “Do whatever it takes to get off the ground” to “Run.” If only there were a place I could go to see what investors were paying for companies with similar metrics to my own?
A decade later, now on the other side of the table as a VC, I find myself in a different position. At Charge Ventures, we look at 000s of early-stage business plans every year. With a much broader network and a decade more experience, my intuition has sharpened significantly. Give me a team, market, and traction, and I can probably get within +-25% of the clearing price. Today I have a different problem.
My job as a VC is to find the “next big thing.” The tricky part is that whatever comes next will look different than whatever came before. The entrepreneurs behind “what’s next” could look and think very differently than those that came before them.
Despite my best efforts to maintain a broad and diverse group of friends and colleagues, my network, like my arteries, is slowly hardening. I am surrounded by people like me, recommending me people like me, raising money for businesses built for me. While one can certainly make money investing in Ivy League software engineers building luxury products for affluent knowledge workers, there is nothing “disruptive” about that plan. It’s the average VC strategy in an industry that loses money on average. It’s retrospective, not forward-looking. It is not the future.
An Argument for Quant VC
So what does the future of VC look like? In addition to being female, I believe it will be much more quantitative. I teach data analytics at Columbia, and am thus clearly biased, but here is the logic:
- Software is eating the world.
- Statistics (i.e. Machine Learning) is eating software.
- Industries are being upended by data-driven businesses lead by data-driven entrepreneurs.
- And yet, the financiers of disruption (VCs), have been slow to adopt data-driven decision making, preferring close networks and qualitative measures.
- Why? The typical refrain is that nascent startups lack the metrics necessary for predictive analytics.
- This is changing though. A new breed of VCs such as Tribe, Correlation, and .406 ventures are leveraging data with interesting results.
- While pre-seed investing will always have a large qualitative component, at Charge we believe that even early-stage investing will be transformed by data-driven technology.
- As private markets mature and become increasingly competitive, firms will look to technology for an edge in the form of web-scale tools for sourcing, scoring, and evaluating entrepreneurs and their ideas.
- Unable to meet face to face, COVID will accelerate this trend and increase the relative importance of quantitive evaluation.
- This makes it all the more important for entrepreneurs to understand their KPIs and how VCs evaluate them.
- On plus side, this should lead to less biased decision making.
- On the other hand, it also creates potential dangers in the form of AI bias.
- Eager to predict the next Mark Zuckerberg or Elon Musk, VCs will use historical data to build their models.
- Similar to flawed models for identifying candidates and criminals, without careful construction, venture capital models will reflect social inequities and biased human decisions of the past.
- Thus, as VCs become increasingly reliant on their algorithms, entrepreneurs must become increasingly aware of how those algorithms work.
A Modest Proposal for Less Bias and Better Outcomes for all
“If I had an hour to solve a problem I’d spend 55 minutes thinking about the problem and 5 minutes thinking about solutions” — Einstein
The best solutions derive their form from the problems they solve. Charge has historical company information but seeks novel ideas from diverse founders. Founders, especially first-timers from non-traditional backgrounds, seek better benchmarks around traction and valuation to evaluate and promote their progress. The future for both groups is paved with data.
In an effort to connect the dots, I’m happy to announce the humble beginnings of our new initiative: “Stats for Startups.”
What
Stats for Startups seeks to quantitatively explore, illuminate, and hopefully combat bias in early-stage venture capital financing. Both that of the industry at large and our own. This is a big task so we will start small with a simple goal: Better Benchmarking for Startup KPIs. Via a bi-weekly series of blog posts, we will explore KPIs like Customer Acquisition Cost (CAC), Average Order Value (AOV), and Payback Period across a variety of industries.
Why
Venture capital decision making is notoriously opaque. As VC becomes metrics-driven, it is more important than ever for Entrepreneurs to understand their KPIs and how VCs evaluate them. This is especially important for new entrepreneurs from diverse backgrounds who often lack access to warm introductions and whisper numbers. KPIs are particularly hard to come by for very early-stage startups due to their stealthy nature, high mortality rate, and focus on action vs. observation.
How
To build our benchmarks, we are anonymously aggregating, normalizing, and analyzing KPIs from thousands of startups decks across the web and our archives. Most large funds already do this to some extent, they just don’t share with everyone else. We will.
When
Our goal is to write a data-driven post every other week, published to this account.
Who
I’m incredibly proud of the team we’ve assembled, though we are actively looking to fill it out:
Alessandro Piazza, Professor of Strategic Management at Rice University, is my Co-Principal Investigator (think co-founder but for research). We met at Columbia Business School, where Alessandro got his PhD studying stigma and scandals across organizations, social movements, status and networks, the determinants of success in creative industries, and the role of sociological mechanisms in angel investing.
Thanos Papadimitriou, Adjunct Professor of Information, Operations, and Management Sciences at NYU Stern and Partner at Charge Ventures. A practicing computer scientist, Thanos has spent his career straddling entrepreneurship, VC, and Academia. We met over a decade ago when he advised my Fulbright research into startup resource (mis)allocation.
Thomas Mecattaf, Data Scientist. Thomas recently received his BS and MS in operations research from Columbia, where I first met him as a student in my analytics class.
In addition to the day to day team, this project would not be possible without blood and sweat of our founding data interns Brittany Wright and Grant Gutstein.
FAQ
Why do you have so many white guys on your team if your goal is to research bias?
Touche. We are acutely aware of this problem. We are doing this is because we know that our networks are narrow and we are therefore biased. We want to become better investors by broadening our networks and better understanding our own biases. One of the objectives of this blog post is to get the word out to potential collaborators. Hit me up: brett at charge dot vc.
What’s in it for Charge?
This is not a charity project. As noted above, we think this work will help us make better investments. That said, we do think better benchmarks and a conversation around bias is a positive contribution to the startup community. We are also data-loving nerds who think KPIs are inherently cool.
Who is paying for this?
Stats for startups is made possible through generous support from the Ewing Marion Kauffman Foundation for Entrepreneurial Research. We received our grant as part of the Knowledge Challenge.
What’s the difference between what you are building and existing startup databases like Crunchbase?
Stats for Startups is focused on 1) lower-level operational KPIs like CAC and CLTV and 2) trying to understand the role of diversity and potential biases in VC decision making. In fact, our work wouldn’t be possible without the awesome folks at Crunchbase, who have generously shared their expertise and data with us to help us make sense of our own (thanks team Crunchbase!). Finally, because we are funded Kauffman, we don’t have to worry about monetizing our work, which means that everything we create will always be available to entrepreneurs for free.
What about privacy?
Stats for Startups will only post normalized, aggregated, and anonymous information. As a VC, it’s mission-critical for us to maintain the trust of entrepreneurs. As such, we maintain every piece of information we come across in the strictest confidence. Nothing about that changes.
What might go wrong?
Lots of things! The vast majority of data on very early-stage companies is currently unstructured, trapped in PDFs and call notes. Even if we can get it out, the data might be too sparse to draw meaningful conclusions. Interpreting has plenty of its own challenges, and if we aren’t careful, we could introduce our own bias into the mix.
What’s next?
This is just the first in what will be a series of data-driven blog posts on bias in VC. Follow Charge on Medium, Linkedin, or our new Twitter to stay tuned!
